Hi,
currently I'm trying to reproduce your results and for the evaluation of the Spring dataset I was wondering which data you used for the validation split since Spring only provides a train-test split. For me the python evaluate.py --cfg config/eval/spring-M.json --model models/Tartan-C-T-TSKH-spring540x960-M.pth from your eval.sh script fails with ValueError: Spring val directory does not exist: datasets/spring/val.
Hi, currently I'm trying to reproduce your results and for the evaluation of the Spring dataset I was wondering which data you used for the validation split since Spring only provides a train-test split. For me the
python evaluate.py --cfg config/eval/spring-M.json --model models/Tartan-C-T-TSKH-spring540x960-M.pth
from youreval.sh
script fails withValueError: Spring val directory does not exist: datasets/spring/val
.